Measurement and Determinants of Supply Chain … and Determinants of Supply Chain Collaboration CIRRELT-2008-17 1 overviewed, followed by a review of the literature dealing with the

___________________________ Measurement and Determinants of Supply Chain Collaboration

Égide Karuranga Jean-Marc Frayret Sophie D’Amours

May 2008

CIRRELT-2008-17

Measurement and Determinants of Supply Chain Collaboration

Égide Karuranga1, Jean-Marc Frayret2,3,*, Sophie D’Amours1,2

1 Département de génie mécanique, Pavillon Adrien-Pouliot, Université Laval, Québec, Canada G1K 7P4

2 Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation (CIRRELT)

3 Département de mathématiques et génie industriel, École Polytechnique de Montréal, C.P. 6079, succursale Centre-ville, Montréal, Canada H3C 3A7

Abstract. Supply chain management and inter-firm collaboration is a growing field of

study for the researcher and a hot topic for the supply chain manager. Despite this

unanimous interest, the measurement of supply chain collaboration has received little

attention to date. This paper proposes a reliable and valid measurement model of supply

chain collaboration based on data from 321 companies from the Quebec forest products

industry using structural equations modeling. In spite of the general low level of

collaboration in this traditional industry, a first order measurement model followed by a

second order confirmatory factor analysis suggest that collaboration in the forest products

industry can be defined by four indicators shared by both customer and supplier sides and

2 indicators specific to each side.

Keywords. Collaboration, supply chain, structural equation model, forest products

industry.

Acknowledgements. This work was funded by the FOR@C Research Consortium in E-

Business in the Forest Products Industry. The authors would like to thanks the anonymous

reviewers of previous versions of this work.

Results and views expressed in this publication are the sole responsibility of the authors and do not necessarily reflect those of CIRRELT. Les résultats et opinions contenus dans cette publication ne reflètent pas nécessairement la position du CIRRELT et n'engagent pas sa responsabilité. _____________________________

* Corresponding author: [email protected] Dépôt légal – Bibliothèque nationale du Québec, Bibliothèque nationale du Canada, 2008

© Copyright Karuranga, Frayret, D’Amours and CIRRELT, 2008

1 Introduction

Supply chain management and inter-firm collaboration is a growing field of study for the

researcher and a hot topic for the supply chain manager. The former sees an opportunity to

develop new models and technology to support collaborative practices, while the latter sees an

opportunity to reap the benefit of supply chain management. Despite this unanimous interest, the

measurement of supply chain collaboration has received little attention to date (Sanders and

Premus, 2005; Jap, 1999; Injazz and Paulraj, 2004). This paper seeks to fill this gap by providing

a reliable and valid measurement model of collaboration intensity and its determinants.

The paper is organized as follows. Section 2 presents a review of the literature concerning the

practices of supply chain collaboration and its measurement methods identified in previous

studies. Next, the research questions addressed in this study are stated. The questionnaire

development and the data collection procedures are described in the fourth section. Then, data

analysis methodologies are presented in the fifth section. The sixth section presents the main

results showing twelve reliable and valid indicators of the collaboration level. The last section

discusses the implications for researchers and practitioners.

2 Literature review

Simatupang and Sridharan (2002) define supply chain collaboration as being “two or more chain

members working together to create a competitive advantage through sharing information,

making joint decisions, and sharing benefits which result from greater profitability of satisfying

end customer needs than acting alone.”

Supply chain collaboration is widely studied from different perspectives. This review of the

literature first proposes an overview of various classification schemes of supply chain

collaboration. Next, the literature dealing with the measure of supply chain collaboration levels is


CIRRELT-2008-17 1

overviewed, followed by a review of the literature dealing with the study of the determinants of

collaboration. Finally, the role of IT in collaboration is also introduced.

2.1 Classification schemes of collaboration

Traditionally, transaction-oriented relationships between organizations have led companies

within supply chains to act as competitors trying to take advantage of each other to maximize

one’s own utility. In such a context, companies see themselves as buyers of goods and sellers of

value-added goods. This adversarial business model promotes opportunistic behaviors as the

profit made by others is seen as an opportunity to improve one’s own benefits.

With the progressive adoption of modern supply chain practices (e.g., VMI, CPFR, early supplier

involvement), several authors have proposed evolution models of B2B relationships, from purely

transactional to collaborative. These relationships models are generally based on a three-level

scale from short-term transactional exchanges to long-term relationship involving several

dimensions of collaborative practice including collaborative product design and joint investment

(Merli, 1991; Lapide, 2002). The development of such frameworks generally considers that B2B

relationships evolve from transactional through various stages to collaboration, which leads to

classification schemes of B2B relationships that provide qualitative measurement scales of

supply chain collaboration levels.

Along this line, Poulin et al. (1994) propose a detailed classification of inter-organizational

relationships on a continuum from the traditional supplier relationship to the creation of co-

enterprises. The study of these relationships leads the authors to the identification of three generic

classes from adversarial to collaborative relationships: supply relationships, outsourcing

relationships, and co-production relationships.


CIRRELT-2008-17 2

Others classification schemes of collaboration have similarly been proposed in the literature.

However, instead of identifying classes of practices, these schemes propose to classify the means

used by companies to introduce collaboration within their organization. For instance, Simatupang

and Sridharan (2002) identify six means of collaboration, which include the setup of joint

objectives to better control demand variability; the integration of decision rules to improve

response coordination to market signal; the use of appropriate measures of performance to align

the partners’ efforts; the creation of coherent decision domains to improve supply chain

decisions; information sharing to improve decision making; and the use of an appropriate

incentive system to improve supply chain decisions. Similarly, Frayret (2002) identifies six

generic patterns of collaboration (see Table 1).

Table 1. Patterns of collaboration (Frayret (2002))

Type Principle Examples of practice

Outsourced local decision-making

Increase the level of responsibility of suppliers

Vendor-managed inventory, supplier quality program, early supplier involvement (component design)

Improved local decision-making

Enhance and align local decision-making

Information sharing, shared POS data, joint capacity management, joint inventory management

Decision objective alignment

Joint objective planning and objective alignment

Collaborative forecasting, collaborative promotion planning, early supplier involvement, category management

Pooled resource and capacity sharing

Resource pooling and sharing, and joint investment

Shared pallets, joint trailer, 3PL (mediated resource sharing), joint facility and R&D investment, shared prototyping facility

Process and IS integration

Business process and information system integration

JIT supplier (ex.: Toyota), CPFR standards adoption, B2B marketplace, E-business standards compliant

Supply chain process reengineering

Internal business process redesign and alignment

Joint cycle time reduction, supply chain event management, suppliers training and evaluation, process postponement, performance metrics


CIRRELT-2008-17 3

These patterns of collaboration include the outsourcing of decision rights; the coordination of

supply chain decisions; the identification of common goals, which indirectly tends to coordinate

supply chain decisions; the pooling of resources; the use of transactional integration framework

and technologies; and the undertaking of a multi-company effort of supply chain reengineering.

Differently, some authors also propose to classify the opportunities of supply chain collaboration.

For instance, Lapide (2002) identifies three areas of opportunities: between a manufacturer and

its suppliers, between a manufacturer and its distribution channel members, and between the

manufacturers and a logistic service provider. Frayret et al. (2003) present an activity-based

synthesis of a wider range of opportunities by studying collaborative practices involving research

and development, procurement, production, marketing and distribution and sales. These

opportunities can also be classified according to the nature of the partner involved (Figure 1).

Figure 1: Collaboration Opportunities (adapted from Frayret et al. 2003)


CIRRELT-2008-17 4

2.2 Measures of supply chain collaboration level

Past and recent studies on collaboration have taken different paths in measuring collaboration

levels. Certainly the first approach to assess the level of collaboration between supply chain

partners is based on the evolution models of B2B relationships mentioned in the previous section.

This type of approaches gives a qualitative assessment of the collaboration level of a relationship

by identifying the closest class of collaboration level it belongs to.

Other authors have studied the impact of particular B2B collaborative practices on supply chain

performance. By doing so, they propose empirical assessment methods of collaboration levels

through five-point scales designed to evaluate the adoption level of a set of specific collaborative

practices. For instance, Stank et al., (2001) develop a measurement method that considers both

internal and external collaborative practices. Similarly, Vereecke and Muylle (2006) consider in

their method both relationships with key suppliers and customers.

Simatupang and Sridharan (2005) propose a similar approach, although more detailed, to measure

a quantitative index of collaboration using a three-dimensional framework based on information

sharing, decisions synchronization and incentive alignment. Each of these dimensions includes

several items specifically developed for retailer-supplier collaborative practices. The evaluation

of the level of adoption of each practice (using also five-point scales) contributes to measuring

the level of information sharing, decisions synchronization and incentive alignment practices, and

ultimately the collaboration index.

2.3 Collaboration: a multiple causes\multiple indicators construct?

Similarly, several paths have been explored in the literature to measure and study the

determinants of collaboration. For instance, a strong stream of research based on characteristics


CIRRELT-2008-17 5

of actors and type of collaboration is represented, among others, by Scott (1991), Wasserman and

Faust (1997), and Fritsch and Lukas (2001). For these authors, operationalized characteristics

include institutional background, resources, psychological and intellectual predispositions, and

learning styles. Large firms are found to be more apt to collaborate than small ones, especially in

research and development activities. In Kleinknecht and Van Reijnen (1992) and Fritsch and

Lukas (2001), the effects of the type of partner and type of agreement are studied as well.

In Heide and John (1992), the authors study a slightly different context characterized by strong

buyers facing small suppliers with the risk of an opportunistic behavior endangering the expected

collaboration. They find that norms play a great role in leveraging obstacles. In contrast, the

dynamic of strong suppliers with a myriad of small buyers was investigated by Berthon et al.

(2003). Converse to Heide and John’s findings, Berthon et al. (2003) empirically find that

relational norms has no effect as governance mechanisms safeguarded against opportunistic

behavior in the presence of transaction-specific assets. However, power dependence theory has

been validated as the size of the buyer positively influenced buyer control.

According to Achrol (1997), collaborators in a network are also described by their density, their

multiplicity, their reciprocity of ties, and their shared values. Network density has also been

studied by Farrell and Klemper (2004) with an emphasis on “instrumental intimacy” in which

firms feed each other through reciprocal influences in the design of new products and application

of business strategies. Trust and commitment are antecedents and central characteristics of

collaboration. The involvement of a supplier in the development of a new product can be rated in

different levels. For instance, the highest level of collaboration occurs when there is an exchange

of staff in the design, development and testing of new products and ideas (Handfiel et al. 1999;

Wasti and Liker, 1997). In Riiter et al. (2002), the authors identify four main antecedents with


CIRRELT-2008-17 6

strong impacts on a company’s network competence: access to resources, network orientation of

human resource management, integration of intra-organizational communication, and openness of

corporate culture. Similarly, Morten and Nohria (2004) identify four barriers to inter-unit

collaboration: unwillingness to seek input and learn from others, inability to seek and find

expertise, unwillingness to help, inability to work together.

2.4 Role of IT in collaboration

Clark, Croson and Schiano (2001) suggest a seven-stage model from independence to virtual

integration while studying regularities occurring in technology-enabled inter-organizational

relationships. Their seven levels of organizational interconnectivity include from bottom to top:

physical data transfer, technology-supported document transmission, electronic data interchange,

new information-intensive process and data sharing, new policies and integrated operations, joint

optimization relationships, virtual channel integration.

In their seven stage model, information technology is believed to enhance collaboration levels but

only under certain conditions. However, Lee, Pak, and Lee (2003) found that “many suppliers

and buyers who implemented B2B networks are still unwilling to cooperate or share data

because of the fear that such information sharing could weaken their negotiating position with

their channel partners.” These authors explain that maximum payoffs are only realized when

information is shared adequately, which is inevitably linked to an increasing interdependency due

to shared IT and business infrastructure. Similarly, the business process has to be reshaped so that

IT can help increase the collaboration level. Similar concerns were raised by Teng et al. (1994)


CIRRELT-2008-17 7

3 Research questions and objectives

In this paper, we propose to study B2B practices and collaboration determinants throughout

several echelon of the Québec forest products supply chain, which includes sawmills (timber first

transformation), paper and cartons producers, furniture manufacturers and other value-added and

building material producers. This study is thus not limited to retailer-supplier relationship as in

Simatupang and Sridharan (2005). Furthermore, as proposed by Vereecke and Muylle (2006),

this study involves the systematic analysis of both supplier and customer sides. The concept of

collaboration is thus here not considered as being a construct describing the intricacy of business

relationships; it is rather considered as an organizational construct, which qualifies the ability of

an organization to develop collaborative ties with its supply chain partners. This perspective of

collaboration can ultimately apply to an industry sector to characterize its B2B practices. This

study proposes to do so in the Québec forest product industry.

3.1 Research questions

In this paper, three research questions are drawn from the above literature review:

Can we determine a collaboration level and its underlying factors?

Does this collaboration level significantly vary across the categories of the value chain?

Do IT, age of companies, their size, internationalization, and position on the value chain play

a key role in enhancing the collaboration level?

3.2 Objectives

The collaboration level as a construct with sound psychometric properties has been rarely

measured from either the customer or supplier perspective. The first objective of this paper is to


CIRRELT-2008-17 8

fill this gap. This study also proposes to contribute to the underlying factors of the collaboration

level and its variation within an industry. A third objective is to assess the role played by

information technology on the adoption of collaborative practices. While there is an extensive

literature related to IT adoption, uses, perceptions, barriers and projects in the North American

forest industry (Karuranga et al., 2005; Vlosky, 2001; Dupuy and Vlosky, 2000), no research has

been devoted to the study of the impact it exerts on collaborative practices.

4 Research design and methodology

4.1 Questionnaire development and Data collection

In order to meet our objectives, we conducted a survey of forest products companies in the

province of Québec, Canada. Our questionnaire was based on the abovementioned research

questions, our objectives, and the existing literature related to the measurement of the

collaboration level with its underlying factors. It was pre-tested with different forest industry

professionals in Quebec. Company leaders were asked to separately assess their collaboration

practices from both customer and supplier sides. The pretest phase yielded comments that

enabled us to refine the final instrument in appendix A. This final instrument was distributed to

695 forest products companies in the province of Quebec. Valid responses were received from

top executives of 312 companies. The 44.9% response rate is far beyond the commonly accepted

standard in such surveys which usually ranges from 15 to 35%. In order to enhance the reliability

of the 312 received valid responses, additional data and cross-validation information were

gathered from public and professional databases including, the Canadian Ministry of Industry,

Statistique Québec and I-CRIQ database.


CIRRELT-2008-17 9

4.2 Profile of respondents

The profile of respondents is presented in Figure 2 by age (i.e., year of creation) and position in

the supply chain. As mentioned earlier, the forest products supply chain includes sawmills, pulp

and paper mills, furniture manufacturers, and other value-added building materials, which

includes doors and windows manufacturers.

Figure 3: Profile of Respondents (sector and year of establishment)

4.3 Non-response bias test

Non-response bias test methods typically include a comparison of the distribution of respondents’

and non-respondents’ characteristics if and when they are known. Similar independent surveys

can also serve as a base of comparison for selected characteristics. In this paper, data from the I-

CRIQ database were used to compare respondents (n = 145) and non-respondents (n = 139) in the

sawmill category, which is the largest group in our survey. A two-sample t-test, performed after

verifying that the samples were similar in shape, showed no statistically significant response bias

between the respondents and the non-respondents with regard to number of employees and year

of establishment.


CIRRELT-2008-17 10

5 Data Analysis

5.1 Indices of collaboration

The collaboration level of each respondent is assessed in this study with the questionnaire

provided in Appendix A. The B2B practices of both supplier and customer sides were used to

assess collaboration levels. Instead of using a five-point scale, respondent were asked to state in

“yes” or “no” answers what type of collaboration they have initiated with their suppliers and

customers. Computed scores of “yes” answers show an evolutionary collaborative path with both

customers and suppliers. This “yes” or “no” response approach is similar to the one proposed by

Link and Bauer (1987) who used a dummy variable of the value 1 if the firm cooperate and the

value 0 if the firm does not. To further improve the robustness of the construct, every “yes”

response was assigned a value on a scale ranging from 0 to 4 referring to the position on the

continuum. The scale itself is built from the aggregation of responses for each B2B practices

where we assume that the higher the collaboration level, the less the number of “yes.” In other

words, this approach assumes that collaborative practices are the most advanced B2B practices.

Therefore, in a traditional industry sector such as the forest product industry, they are less likely

to have been adopted by most companies, only by the most advanced of them.

5.2 Measurement model

In order to respond to the first research question, we developed a measurement model following

two steps: an exploratory factor analysis and then a confirmatory factor analysis. The obtained

results were used to respond to the second and the third research questions through a rigorous

purification process. In the absence of an a priori theoretical model, we followed Schumacker

and Lomax‘s (2004) recommendation and ran both exploratory and confirmatory factor analysis.


CIRRELT-2008-17 11

5.3 Exploratory factor analysis

The following constraining rules were applied to our exploratory factor analysis: Kaiser-Meyer-

Olkin Measure of Sampling Adequacy must be bigger than 0.80, communalities extraction must

be bigger than 0.50, variance explained must be bigger than 0.50, factor loading must be bigger

than 0.50, double loading must be bigger than 0.50, and single loading for sake of parsimony.

Varimax rotation with Kaiser normalization was applied with Eigen-values larger than 1. We

calculated a correlation matrix in a search for multicollinearity. Internal reliability of our

constructs was assessed using Cronbach’s alpha coefficient bigger than 0.70.

5.4 Confirmatory factor analysis

In order to confirm the factor structure generated at the exploratory level, a confirmatory factor

analysis was conducted following Byrne’s (1994) recommendations. To avoid measurement

biases while assessing the degree to which the data fit the model, Browne and Cudeck (1993) and

Hu and Bentler (1999) recommend numerous fit statistics structural equations models. We used

the most cited indices to assess the fitness of our model to the 312 companies in the survey:

- the ratio of chi-square to degree of freedom (χ2/df): an acceptable range is met between 2

and 3 (Carmines and McIver, 1981). Some conservative schools put the chi-square/df

ratio to less than 2.00 with an insignificant P-value (e.g. P>0.05) (Byrne, 1994). Hartwick

(1994), Hair et al., (1995) and Chau (1997) recommend a maximum of 3.0;

- the root mean-square residual (RMR) which compares the values of variances and

covariances predicted by the model with the variances and covariances obtained from the

data to detect discrepancy between the two, thus zero representing a perfect fit;


CIRRELT-2008-17 12

- the root mean square error of approximation (RMSEA) with value close to 0.06 or better.

Schumacker and Lomax (2004) recommend 0.05 or less for the RMSEA in terms of

providing a good fit of the data;

- the standardized root mean square residual (SRMR) with value close to 0.08 or better (Hu

and Bentler, 1999);

- the comparative fit index (CFI) with value close to 0.95 or better (Hu and Bentler, 1999);

- goodness-of-fit index (GFI) which ranges from zero (no fit) to one (perfect fit).

6 Results and discussions

6.1 Exploratory factor analysis

As shown in Table 1, an exploratory factor analysis performed using SPSS 11.00 yielded three

distinct factors. Customer and supplier items loaded separately with the exception of joint

investments on both sides loading in a separate dimension. A reliability analysis was conducted

for the three factors. As shown in Table 2, deleted items include exchange of performance

evaluation for low extraction. Furthermore, Joint new products development, resource sharing,

and implementation of replenishment systems are suppressed for highly loading on two distinct

factors. Deleted items appear to be those generally consider as being the most advanced forms of

collaborative practices, which indicates a rather poor level of collaboration for the forest product

industry. To further validate the robustness of our instrument, Table 1 presents the Cronbach’s

alfa coefficient. As shown in Table 3, no multicollinearity was detected in the correlation matrix

as all coefficients are inferior to 0.85 with a minimum of 0.10. The output shows a coefficient of

0.81 for supplier side items and 0.80 for those on the buyer side. These rates are higher than the

commonly accepted level of 60% for such an exploratory study (Flynn et al., 1990). Hence, joint

investments both at the supplier and client sides were dropped for poor reliability (0.54).


CIRRELT-2008-17 13

Table 1: Exploratory Factor Analysis

Kaiser-Meyer-Olkin Sampling Adequacy 0,86 6.1.1 6.1.

Rotated Component Matrix. Factor 1 Factor 2 Factor 3

z37e Joint planning 0.74 0.14 0.16

z37g Joint delivery improvement 0.73 0.16 0.11

z37b Exchange of basic information 0.72 0.23 -0.12

z37f Exchange performance evaluation 0.67 0.06 0.30

z37c Resources sharing of logistic assets 0.66 0.12 0.20

z37a Joint sales forecasting 0.63 0.24 -0.07

z35b Exchange of basic information 0.09 0.78 -0.03

z35a Joint sales forecasting 0.14 0.71 -0.06

z35e Joint planning 0.14 0.70 0.15

z35h Joint new products development 0.07 0.68 0.11

z35g Joint delivery improvement 0.26 0.62 0.06

z35d Replenishment systems 0.31 0.58 0.17

z35i Joint investment 0.05 0.20 0.79

z37i Joint investment 0.20 0.00 0.77

Alfa Cronbach’s Coefficient calculated on items in bold 0.81 0.80 0.54

Table 2: Deleted items

Deleted items Level of deletion Reason

z35f Exchange performance evaluation 1st iteration extraction<0.5

z37h Joint new products development 2nd iteration double load.>0.5

z35c Resources sharing of logistic assets 2nd iteration double load.>0.5

z37d Replenishment systems 2nd iteration double load.>0.5

z37j Vendor Managed Inventory 3rd iteration extraction<0.5

z35i Joint investment reliability analysis Cr. alfa <0.7

z37i Joint investment reliability analysis Cr. alfa <0.7


CIRRELT-2008-17 14

Table 3: Multicollinearity Analysis

Inter-items Correlation Matrix z35 a

z35 b

z35 d

z35 e

z35 g

z35 h

z35 i

z37 a

z37 b

Z37 c

z37 e

z37 g

z37 i

z37 f

Collaboration with customers

z35a Joint sales forecasting 1.00

z35b Exchange of basic information 0.51 1.00

z35d Replenishment systems 0.37 0.41 1.00

z35e Joint planning 0.41 0.43 0.42 1.00

z35g Joint delivery improvement 0.35 0.40 0.34 0.43 1.00

z35h Joint new products development 0.33 0.43 0.37 0.37 0.40 1.00

z35i Joint investment 0.11 0.14 0.19 0.22 0.18 0.18 1.00

Collaboration with suppliers

z37a Joint sales forecasting 0.32 0.18 0.30 0.25 0.22 0.19 0.12 1.00

z37b Exchange of basic information 0.26 0.30 0.21 0.25 0.23 0.23 0.10 0.52 1.00

z37c Resources sharing of logistic assets 0.16 0.20 0.36 0.22 0.25 0.10 0.18 0.30 0.41 1.00

z37e Joint planning 0.16 0.16 0.31 0.24 0.33 0.21 0.19 0.41 0.46 0.39 1.00

z37g Joint delivery improvement 0.15 0.20 0.34 0.19 0.38 0.21 0.15 0.35 0.44 0.42 0.55 1.00

z37i Joint investment 0.07 0.05 0.15 0.11 0.06 0.11 0.37 0.15 0.13 0.22 0.24 0.20 1.00

z37f Exchange performance evaluation 0.17 0.13 0.30 0.20 0.22 0.13 0.22 0.31 0.36 0.47 0.46 0.10 0.27 1.00


CIRRELT-2008-17 15

Before submitting the retained items to a confirmatory factor analysis, the univariate distribution

normality test was conducted. In a CFA using maximum likelihood as an estimation method,

acceptable skewness and kurtosis levels for a normal distribution are respectively set bigger than

2 and bigger than 7 (West, Finch, and Curran, 1995). Skeweness and kurtosis coefficients for

items in this study meet the recommended limits.

6.2 Confirmatory factor analysis

Although univariate distribution satisfies conditions for an estimation by the maximum likelihood

method, we use diagonally weighted least square (DWLS) as suggested by Flora and Curran

(2004) and Jöreskog and Sörbom (1996) whenever a stable parameter estimator is needed for

small samples.

We started with separate first order confirmatory models for each side (supplier and customer).

Using LISREL 8.71, the two models were then combined in one measurement model with

covariance between both sides. The combined measurement model is depicted in Figure 3. Next,

a second order confirmatory factor analysis was ran in order to assess the relative contribution of

each side to the collaboration level. Results of second order CFA are shown in Figure 4. All final

models were retained after rigorously checking for residuals and modification indices using a step

by step modification procedure starting by residual matrices followed by modification indices

before deleting any non-significant parameter. The final models fit indices (displayed in Table 4)

exceed the generally accepted standards, thereby confirming that our model fits the data very

well.


CIRRELT-2008-17 16

Figure 4: First Order CFA


CIRRELT-2008-17 17

Figure 5: Second Order CFA


CIRRELT-2008-17 18

Table 4: Measurement Models Fit Indices

Chi2 df Chi2/df P-value RMSEA CFI SRMR GFI

1st order CFA 130.38 53 2.45 0.00000 0.069

2nd order CFA 101.66 53 1.92 0.00005 0.055

Desirable level 2 - 3 0.060 0.95 0.08 0 - 1

6.3 Testing the validity of the model

In order to test the internal validity of a model, structural equations models often rely on content,

convergent, discriminant criteria, and nomological validities. In this study, content and face

validity encompassing wording and phrasing was met at the pretest stage of the questionnaire. An

external validity assessment can be conducted in four different ways: by simulating the results to

meet the real world situation, by testing the model with a new sample, by randomly splitting the

original sample in two new samples to be tested with the same model and by bootstrapping the

results in search of parameter estimates variations (Byrne, 1994; Schumacker et al., 2004)

Firstly, convergent validity is generally achieved when factor loadings are bigger than 0.50

(Kline, 1998). This common rule of loading is met as all items in all the final models show values

greater than 0.50. A second criterion for convergent validity is the variance extracted by every

factor whose recommended threshold is 0.50.

Secondly, to assess discriminant validity, correlations of the factors underlying the set of

indicators were calculated to ensure that constructs can be empirically differentiated from each

other. Limits are usually set between 0.10 and 0.85 (Kline, 1998) for a good discriminant

validity. Our results show that none of the correlation coefficients exceeds the recommended

limits. Moreover, each item loads stronger on its associated factor than on any other factor.


CIRRELT-2008-17 19

Third, to examine criterion validity of the two constructs in the second order factor analysis

model, we estimated a regression model between them as independent variables and the overall

collaboration as a dependant variable. Both constructs are significant and exert positive influence

on the overall collaboration concept as a dependent variable. Fourth, in the absence of previous

studies, nomological validity could not be assessed. It usually determines if an instrument

behaves as it should when compared to empirical literature.

As to external validity aiming at testing the replicability of our results, the two factor model was

randomly split into two samples using SPSS. A first order CFA using LISREL 8.71 and DWLS

as estimating method was run on the two randomly split samples (n1= 156., n2= 156). Results

reported in Table 5 show similar goodness of fit indices in all 3 samples. Compared to the 1st

order CFA model, these fits show no statistically significant difference.

Table 5: External Validity

Fit indice Sample 1 model

N=156 Sample 2 model

N=156 1st order CFA model

N=312

Chi-square 76.26 (P =0.020) 98.73 (P= 0.00014) 130.382 (P = 0.000)

Df 53 53 53 NFI 0.97 0.96 0.975 CFI 0.99 0.98 0.985 GFI 0.99 0.98 0.986

RMSEA 0.053 0.075 0.0685

6.4 Collaboration as a multi item construct

Results obtained from section 6.1 to 6.3 of this paper indicate that collaboration in the forest

products industry can be defined by four items shared by both customer and supplier sides and 2

items specific to each side. The four items which are common practices on both sides are joint

sales forecasting, exchange of basic information, joint planning, and finally, Joint delivery


CIRRELT-2008-17 20

improvement. Two practices are highly oriented towards customers: resources sharing of logistic

assets and exchange of performance evaluation. B2B Practices engaged mostly with suppliers

include the implementation of replenishment systems and Joint new products development.

6.5 Testing invariant structure

Our third research question refers to model validity across groups in the sampled population. To

test this, we conducted an invariant structure analysis across the groups of companies according

to the following criteria using a comparison of correlation matrices: exporting firms versus the

rest, year of establishment to compare young companies established after 1991 to older ones,

position in the supply chain, having a website or not, interacting using IT or not, and number of

employees as a proxy of size. All the correlation matrices were found to be equal, thus the null

hypothesis cannot be rejected. At p-value =0.05, there is no statistically significant difference as

reported in Table 6. This confirms a high level of similarity in the low collaboration intensity

across the forest products industry. More specifically, IT does not seem to influence the

collaboration level in the forest products industry.

Table 6: Invariant Structure Analysis

Chi-Square p-value

Having a website (n=196) versus no website(n=116) 85.123 0.272

Interacting via IT( n=168) versus non interacting (n=144) 75.606 0.556

Export (n=172) versus non export (n=140) 69.062 0.755

Number of employees: 141 large companies versus 171 small ones 77.953 0.480

Domain of activity : 145 sawmills versus 167 other companies 78.238 0.471


CIRRELT-2008-17 21

7 Conclusion, Limitations and Future Research

Collaborative practices are constantly evolving within and between networked companies. The

forces behind this move include IT, industrial culture, size of companies, location, etc. However,

to the best of our knowledge the level of collaboration as such and its underlying factors had

never been measured, at least at the time we conceived and conducted this survey. We developed

a method to fill the gap and stimulate the debate on collaboration construct measurements.

Therefore, the main theoretical contribution of this study remains the development of a

measurement model of collaborative practices in organizations building on the example of the

Quebec forest products industry. Practices are categorized into those which are common to

customers and suppliers and those which are specific to either of these two sides.

In this study, we also focused on relationships within one sector: the forest products industry. The

overall results show a sector rather less prone to collaborating based on indices of collaboration.

There is an apparent strong industrial culture which is likely not in favor of high levels of

collaboration as invariant structure tests across different grouping criteria yield no statistically

significant differences.

As to the practical implications, through the above categorization of collaboration indicators,

managers will find useful indications of how to assess their collaborative practices. Potential

marketers of IT products as well as public institutions will find some key areas that need products

and services to improve collaborative practices in the industry. Hardware and software companies

should also exploit the results if they want to enter the untapped e-supply chain management

market in the Quebec forest industry.

However, despite the above theoretical and practical contributions, some limitations are

discernable. First, the number and quality of interviewed persons could be extended to cover all


CIRRELT-2008-17 22

staff levels of the surveyed organizations. Data should be collected on their personal

characteristics. These data would extend to age, training, previous experience and position in the

organization. The types of tools used to collaborate should also receive more attention in future

studies. As to IT, although weak and with little impact on collaboration levels, its adoption rates

and subsequent collaboration practices enhancement might also have been influenced by other

external actors which potentially include public and private institutions. Investigating ties with

these external organizations could help in predicting the successful implementation of any

innovation at home as Granovetter, (1985) states. Therefore, extending future research to the

level of collaboration with additional actors will certainly bring in more plausible explanations as

to why the forest product industry is lagging behind in both IT and collaboration practices.

Future research should also concentrate on the role played by the history of the forest industry

firms and determine at what extent it is correlated with the low collaboration levels previously

mentioned. Barney (1991) emphasized experience acquisition and accumulation through time.

Past capabilities of changing routines is a key indicator as to how new technology and ideas will

be adopted.

Similarly, an interesting avenue for future research is the comparison of the pre adoption status

and the post adoption results, thus assessing the role of IT impact on collaborative practices. Such

a longitudinal research could also reveal how successful IT tools are implemented at different

companies in order to boost collaboration levels.

The agenda of future research could also include the following issues:

- standardization of practices within the industry and harmonization of channels

- internal cooperation between departments of the organization under study as mentioned

by Kim (2001) and Stank et al. (2001);


CIRRELT-2008-17 23

- commitment, trust, culture, social embeddings of knowledge;

- type of technology used to enhance collaborative practices.

Finally, one of the potential sources of conflicting findings in quantitative research is the reliance

on highly contextual studies which yield non-generalizable results. Using our validated constructs

under different settings will further enhance their validity.

8 Acknowledgement

This work was funded by the FOR@C Research Consortium in E-Business in the Forest Products

Industry. The authors would like to thanks the anonymous reviewers of previous versions of this

work.


CIRRELT-2008-17 24

Appendix A: Survey Questions

Respondents were asked to answer “yes” if they had implemented these practices, “no” if not.

Customers side Suppliers side

Joint sales forecasting (q35a), Exchange of basic information (q35b), Resources sharing of logistic assets (q35c), Replenishment system (q35d), Joint planning (35e), Exchange performance evaluation (q35f), Joint delivery improvement (q35g), Joint new products development (q35h), Joint investment (q35i)

Joint sales forecasting (q37A), Exchange of basic information (q37B), Resources sharing of logistic assets (q37c), Replenishment system (q37d), Joint planning (37e), Exchange performance evaluation (q37f), Joint delivery improvement (q37g), Joint new products development (q37h), Joint investment (q37i), Vendor Managed Inventory (q37j)


CIRRELT-2008-17 25

Literature cited

1. Achrol, R. S. 1997. Changes in the theory of interorganisational relations in marketing:

toward a network paradigm. Journal of the Academy of Marketing Science, 25 (1): 56-71.

2. Barney, J. 1991. Firm resources and sustained competitive advantage. Journal of

Management, 17:99-120.

3. Berthon, P., L. F. Pitt, M. T. Ewing and G.Bakkeland. 2003. Norms and power in marketing

relationships: Alternative theories and empirical evidence. Journal of Business Research, 56

(9): 699-709

4. Browne, M. W., R. Cudeck. 1993. Alternative ways of assessing model fit. In: K.A. Bollen

and J.S. Long, Editors, Testing structural equation models, Sage, Newbury Park, CA, 136–

162.

5. Byrne, B.M. 1994. Structural Equation Modeling with EQS and EQS/Windows: Basic

Concepts, Applications, and Programming. Sage Publications, CA: Thousand Oaks.

6. Carmines, E. G., J.D. McIver 1981. Analyzing models with unobserved variables analysis of

covariance structures. In: G.W. Bohinstedt and E.F.

7. Chau, P. Y. K. 1997. Re-examining a model for evaluating information center success using

a structural equation modeling approach. Decision Sciences, 28(2): 309–334

8. Dupuy, C.A., R.P., Vlosky, 2000. Status of electronic data interchange in the forest products

industry. Forest Products Journal. Madison 50(6): 32-39.

9. Clark, T. H., D. C. Croson and W. A Schiano 2001. Hierarchical Model of Supply-Chain

Integration: Information Sharing and Operational Interdependence in the US Grocery

Channel. Information Technology and Management, 2: 261-288


CIRRELT-2008-17 26

10. Farrell, J., P. Klemperer. 2004. Coordination and lock-in: Competition with switching costs

and network effects. In Handbook of Industrial Organization: Armstrong and R.H. Porter,

Eds. North Holland, Amsterdam: The Netherlands,3

11. Flora, D. B., P. J. Curran 2004. An empirical evaluation of alternative methods of estimation

for confirmatory factor analysis with ordinal data. Psychological Methods, 9:466–491.

12. Flynn, B. B., S. Sakakibara, R. G. Schroeder K. A. Bates and E.J. Flynn. 1990. Empirical

research methods in operations management. Journal of Operations Management, 9(2):250–

284.

13. Frayret, J.M. 2002. A Conceptual Framework to Operate Collaborative Manufacturing

Networks. PhD thesis, Faculté des sciences et de génie, Département de génie mécanique,

Université Laval, Québec, Canada.

14. Frayret, J.M., F., D'Amours, D'Amours, S. 2003. Collaboration et outils collaboratifs pour la

PME manufacturiere. Rapport de recherche du CEFRIO.

15. Fritsch, M., R. Lukas 2001.Who cooperates on R&D? Research Policy, 30(2): 297-312

16. Gimenez, C. 2004. International Journal of Integrated Supply Management (IJISM), 1(1)

17. Granovetter, M. 1985. Economic action and social structure: The problem of embeddedness.

American Journal of Sociology, 91:481-510.

18. Hair, Jr. J. F., R. E.Anderson, R. L., Tatham and W. C. Black. 1995. Multivariate Data

Analysis with Readings. 4th Edition. Prentice Hall. NJ: Englewood Cliffs.

19. Handfiel, R. B., E. L. Nichols. 1999. Introduction to Supply Chain Management. Prentice-

Hall, NJ:

20. Hartwick, J., H. Barki. 1994. Explaining the role of user participation in information system

use. Management Science, 40(4): 440–465.


CIRRELT-2008-17 27

21. Heide, J. B., G. John. 1992. Do Norms Matter in Marketing Relationships? Journal of

Marketing, 56(2): 32-44

22. Hu, L. P. M. Bentler 1999. Cutoff. Criteria for fit indexes in covariance structure analysis:

Conventional criteria versus new alternatives. Structural Equation Modeling , 6: 1–55

23. Injazz, C., J. Paulraj. 2004. A Towards a theory of supply chain management: the constructs

and measurements. Journal of Operations Management, 22(2): 119-150

24. Jap, S.D. 1999. Pie-Expansion Efforts: Collaboration Processes in Buyer-Supplier

Relationships. Journal of Marketing Research, 36(4): 461-475

25. Jöreskog, K. G., D. Sörbom 1996. LISREL 8 user's reference guide. Scientific Software

International. Chicago.

26. Karuranga, E., J.M., Frayret, and S. D'Amours. 2005. E-business in the Quebec forest

industry: use, perception and projects. Journal of Forest Products Business Research.

27. Kim, H. W. 2001. Modeling Inter- and Intra-Organizational Coordination in Electronic

Commerce Deployments. Information Technology and Management, 2:335-354.

28. Kleinknecht, A., J. Reijnen 1992. Why do firms cooperate on R&D? An empirical study.

Research policy, 21(4): 347-360

29. Kline, R. B. 1998. Principles and practice of structural equation modelling. The Guilford

Press. New York, NY.

30. Lapide, L. 2002. Are we moving from buyers and sellers to collaborators? AMR Research

Report.

31. Lee, S. C., B. Y. Pak and H. G. Lee 2003. Electronic Commerce Research and Applications.

Business value of B2B electronic commerce: the critical role of inter-firm collaboration,

2(4): 350-361


CIRRELT-2008-17 28

32. Link, A. N., L.L. Bauer. 1987. Cooperative research in U.S. manufacturing. In: Rothwell, R.,

Bessant, J. Eds., Innovation: Adaptation and Growth. Elsevier, Amsterdam, 147–154

33. Merli, G. 1991. Co-makership: the new supply strategy for manufacturers. Productivity

Press, Cambridge, MA.

34. Morten, T., H. N. Nohria. 2004. How to Build Collaborative Advantage. MIT Sloan

Management Review, Cambridge, 46(1):22

35. Ritter, T., I.F. Wilkinson, W.J. Johnston. 2002. Measuring network competence: some

international evidence. Journal of Business & Industrial Marketing, 17(2/3): 119-138

36. Sanders, N. R. Premus. 2005. Modelling the relationship between firm it capability,

collaboration, and performance. Journal of Business Logistics, 26 (1):1-23

37. Schumacker R. E., R. G. Lomax. 2004. A beginner's guide to structural equation modeling.

New York, NY: Lawrence Erlbaum Associates, 498

38. Scott, J. 1991. Social network analysis: a handbook. Sage Publications, London.

39. Simatupang, M. T., R. Sridharan 2002. The Collaborative Supply Chain. The International

Journal of Logistics Management , 13(1):15-30

40. Simatupang, T.M., R., Sridharan. 2005. The collaboration index: A measure for supply chain

collaboration. International Journal of Physical Distribution and Logistics Management

35(1): 44-62.

41. Stank, T.P., S.B., Keller, P.J. Daugherty. 2001. Supply chain collaboration and logistical

service performance. Journal of Business Logistics 22(1): 29-48.

42. Teng, J. T. C., V.Grove, and K. D. Fiedler. 1994. Re-designing business processes using

information technology. Long Range Planning, 27(1): 95-106


CIRRELT-2008-17 29

43. Vereecke, A., S., Muylle. 2006. Performance improvement through supply chain

collaboration in Europe. International Journal of Operations & Production Management

26(11): 1176 - 1198.

44. Vlosky, R.P. 2001. e-Business in the Forest Products Industries. In Proceedings of the 2nd

World Symposium on Logistics in the Forest Sector, Supply Chain Management for the

Paper and Timber Industries, K. Sjöström, Eds., Växjö, Sweden.

45. Wasserman, S., K. Faust. 1997. Social network analysis: Methods and applications.

Cambridge: Cambridge university press,

46. Wasti, N. L., J. K. Risky. 1997. Business or competitive power? Supplier involvement in

Japanese product design. Journal of Product Innovation Management, 14(5): 337-355

47. West, S. G., J. F. Finch, and P. J. Curran. 1995. Structural equation models with nonnormal

variables problems and remedies. In: R.H. Hoyle, Editors, Structural equation modeling

concepts, issues and applications. Sage, CA: Newbury Park,, 56–75


CIRRELT-2008-17 30

Documents

Measurement and Determinants of Supply Chain … and Determinants of Supply Chain Collaboration CIRRELT-2008-17 1 overviewed, followed by a review of the literature dealing with the